Claude Agent Skills

Author a new, well-formed Claude Code skill, agent, or hook from a one-line intent. Decides the right artifact type (skill vs agent vs hook) and placement (global vs project), then writes a correct SKILL.md / agent file / settings.json hook with engineered trigger phrasing, a clear procedure, and guardrails. Global and project-agnostic. Trigger when the user says "new skill", "new-skill", "create a skill", "make an agent", "scaffold a skill/agent/hook", or "author a Claude command".

Scans the current project's .agents/skills/ and .codex/agents/ directories for items missing from the project's .claude/skills/ and .claude/agents/ directories, presents the gap list for approval, adapts each approved item to Claude Code format, and writes to the project-local .claude/ directory. Never writes to global ~/.claude/ or modifies .agents/ or .codex/. Trigger when the user says "port project claude", "port-project-claude", "replicate project codex skills for claude", "sync project ...

Scan the current repository and recommend which directories deserve their own nested CLAUDE.md file (i.e. where to run /init), as a ranked report with a one-line rationale per directory. Global and project-agnostic. Does NOT generate the files - it tells you where /init is worth running. Trigger when the user says "where should I run /init", "which directories need a CLAUDE.md", "where-claude", "scout claude.md placement", or "find nested CLAUDE.md candidates".

Audit Codex AGENTS.md and AGENTS.override.md instruction files in the current project for staleness and coverage gaps. Scores freshness with git churn, flags content drift, discovers directories that may deserve scoped AGENTS.md guidance, and reports ranked findings with per-file approval before edits. Global and project-agnostic. Trigger when the user says "audit AGENTS.md", "audit agent docs", "audit-agent-docs", "are my AGENTS.md files up to date", "check if AGENTS.md is stale", "Codex doc...

Implement persistent memory patterns for AI agents using AgentDB. Includes session memory, long-term storage, pattern learning, and context management. Use when building stateful agents, chat systems, or intelligent assistants.

Audit and maintain an existing Obsidian Vault against the current state of the repository it covers. Runs two parallel read-only scans (codebase + vault), produces a structured audit report across 11 checks (missing coverage, stale notes, dead wikilinks, orphan notes, tag drift, frontmatter drift, ADR staleness, canvas freshness, _AI-CONTEXT.md drift, Dataview query health, plugin recommendations), then applies fixes only after per-item user approval. Global and project-agnostic; pairs with t...

Generate a comprehensive, ground-truth-anchored Obsidian Vault for any repository. The vault covers codebase understanding, execution planning, Architecture Decision Records (ADRs), session reasoning logs, and design tracking — structured for both human navigation and AI consumption. Creates the full folder skeleton, note templates, .obsidian/ config (plugin manifest, graph settings, CSS color snippets), tag taxonomy, AI context file, and a starter architecture canvas. Idempotent: detects an ...

Periodic repo health check orchestrator. Runs four maintenance skills in sequence — CI status check (/diagnose-ci), CLAUDE.md staleness audit (/audit-claude-md), documentation accuracy audit (/audit-docs), and Obsidian Vault sync (/obsidian-vault-sync if a vault exists) — pausing for approval between steps and producing a single consolidated findings report at the end. Global and project-agnostic. Trigger when the user says "health check", "health-check", "repo health", "check repo health", "...

Spawn Claude sub-agents to debate a topic, idea, proposal, or decision from multiple perspectives. A Proponent argues for it, a Critic argues against, an optional Devil's Advocate challenges both, and a Synthesizer produces a structured verdict. Use when you want rigorous multi-perspective analysis rather than a one-sided answer. Global and project-agnostic. Trigger when the user says "debate this", "debate idea", "argue both sides", "steelman and critique", "agents debate", "debate topic", "...

Translates current-project Codex permission rules (.codex/rules/default.rules prefix_rule blocks only, not personal ~/.codex rules) into Claude Code permission entries (.claude/settings.local.json), computes the gap against already-present rules, and presents portable rules for per-rule approval grouped by category before writing. Global and project-agnostic. Trigger when the user says "sync claude permissions", "sync-claude-permissions", "port project codex rules to claude", "translate proje...

Security matters because skills can shape agent behavior, suggest commands, and include supporting files. Start with high-grade skills and review source before installation.

Create a SemVer annotated git tag on main by inferring the bump type (major/minor/patch) from conventional commits since the last tag, push the tag to remote, and optionally create a GitHub Release via gh. Global and project-agnostic. Trigger when the user says "tag main", "create a tag", "tag release", "cut a release", "release v1.2.3", "tag version", "bump version and tag", "create a release", "release this", or "/tag". SKIP when the user is asking about Docker image tags or non-git tagging.

Scans the current repo for CLAUDE.md files, finds directories that lack a matching AGENTS.md, translates each CLAUDE.md to an AGENTS.md draft (applying the Claude -> Codex substitution table), runs a $where-agents-md sanity check per candidate, and - with per-item approval - either writes the translated draft or invokes $init inline for a richer code-derived doc. Never overwrites an existing AGENTS.md. Global and project-agnostic. Trigger when the user says "claude md to agents md", "bootstra...

Scans the current project's .claude/skills/ and .claude/agents/ directories for items missing from the project's .agents/skills/ and .codex/agents/ directories, presents the gap list for approval, adapts each approved item to Codex format, and writes to project-local Codex paths. Never writes to global ~/.codex/ or modifies .claude/. Global and project-agnostic. Trigger when the user says "port project codex", "port-project-codex", "replicate project claude skills for codex", "sync project .c...

Translates current-project Claude Code permission rules (.claude/settings.local.json permissions.allow and permissions.deny Bash entries only) into Codex prefix_rule blocks in .codex/rules/default.rules, computes the gap against existing project-local Codex rules, and presents portable rules for per-rule approval before writing. Global and project-agnostic. Trigger when the user says "sync codex permissions", "sync-codex-permissions", "port project claude permissions to codex", "translate pro...

Create or update a GitHub Pull Request from the current branch. Reads all commits since the base branch, auto-generates a structured PR title and body (Summary / Changes / Test Plan sections), infers labels from commit types, and shows everything for approval before running gh pr create. Handles push if needed and detects existing PRs. Global and project-agnostic. Trigger when the user says "create a PR", "open a pull request", "make a PR", "/pr", "submit PR", or "push and open a PR".

Scan the current repository and recommend which directories deserve their own scoped Codex AGENTS.md or AGENTS.override.md instruction file, as a ranked report with one-line rationales. Does not generate files. Trigger when the user says "where should AGENTS.md go", "which directories need AGENTS.md", "where-agents-md", "scout AGENTS.md placement", "find nested AGENTS.md candidates", or "where should Codex instructions live".

Scan the current repository, infer its workflow, and propose a short, ranked, high-ROI plan of Claude Code skills, agents, and hooks tailored to it - then scaffold only the items the user approves. Global and project-agnostic. Trigger when the user says "what skills should this project have", "propose automation", "set up skills for this repo", "scaffold agents/hooks", or "plan automation".

Diagnose and fix GitHub Actions CI failures for the current repository. Fetches failing job logs via gh CLI, reproduces the failure locally, identifies the root cause (code bug, flaky test, or infra/config issue), proposes a targeted fix, applies it after approval, then offers to commit and push to re-trigger CI. Global and project-agnostic (requires gh CLI). Trigger when the user says "diagnose CI", "diagnose-ci", "why is CI failing", "CI is red", "fix CI failures", "debug GitHub Actions", "...

Use when performing code review, writing or refactoring code, or discussing architecture; complements clean-code and does not replace project linter/formatter.

Analyze a GitHub repository and propose or improve a GitHub Actions CI setup with clearly separated jobs — one job per concern — so each check shows independently in the GitHub PR and Actions UI. Global and project-agnostic. Trigger when the user says "set up CI", "setup-ci", "add GitHub Actions", "propose CI jobs", "improve my CI", "add CI pipeline", "set up GitHub Actions workflow", "create GitHub Actions tests", or "suggest CI checks". SKIP when the user is asking about a non-GitHub CI sys...

A custom AI agent skill for AWS Infrastructure.

A custom AI agent skill for Rate Limiting & DDoS Prevention.

A custom AI agent skill for Next.js Image Optimization Expert.